Variational Mixture of Graph Neural Experts for Alzheimer's Disease Biomarker Recognition in EEG Brain Networks

2025-10-13

AI Summary

Alzheimer’s disease (AD) and frontotemporal dementia (FTD) exhibit substantial electrophysiological overlap in electroencephalography (EEG), impeding reliable subtype differentiation and severity staging. To address this, we propose a frequency-specific graph neural network framework: it integrates multi-granularity Transformers to extract time-series features from δ, θ, α, and β bands; employs a variational graph convolutional encoder regularized by a Gaussian Markov random field prior; and incorporates an adaptive gating mixture-of-experts mechanism that explicitly links band-specific neurophysiological interpretations to model decisions. This structured variational inference approach balances interpretability and discriminative power. Evaluated on two independent EEG cohorts, our method achieves 4–10% AUC improvements over state-of-the-art methods in both subtype classification and clinical staging. Furthermore, learned expert weights show significant correlations with established clinical rating scales, supporting the discovery of novel, frequency-resolved biomarkers for neurodegenerative dementias.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Models for Web evolution

Abstract

Dementia disorders such as Alzheimer's disease (AD) and frontotemporal dementia (FTD) exhibit overlapping electrophysiological signatures in EEG that challenge accurate diagnosis. Existing EEG-based methods are limited by full-band frequency analysis that hinders precise differentiation of dementia subtypes and severity stages. We propose a variational mixture of graph neural experts (VMoGE) that integrates frequency-specific biomarker identification with structured variational inference for enhanced dementia diagnosis and staging. VMoGE employs a multi-granularity transformer to extract multi-scale temporal patterns across four frequency bands, followed by a variational graph convolutional encoder using Gaussian Markov Random Field priors. Through structured variational inference and adaptive gating, VMoGE links neural specialization to physiologically meaningful EEG frequency bands. Evaluated on two diverse datasets for both subtype classification and severity staging, VMoGE achieves superior performance with AUC improvements of +4% to +10% over state-of-the-art methods. Moreover, VMoGE provides interpretable insights through expert weights that correlate with clinical indicators and spatial patterns aligned with neuropathological signatures, facilitating EEG biomarker discovery for comprehensive dementia diagnosis and monitoring.

Problem

• Identifying overlapping EEG biomarkers for Alzheimer's and frontotemporal dementia
• Improving dementia subtype differentiation and severity staging accuracy
• Overcoming limitations of full-band frequency analysis in EEG diagnosis

Innovation

• Variational mixture of graph neural experts
• Multi-granularity transformer for frequency bands
• Structured variational inference with adaptive gating